Determining the value of the fine-structure constant from a current balance: getting acquainted with some upcoming changes to the SI Richard S. Davis* International Bureau of Weights and Measures (BIPM), Sèvres, France 92312 Abstract The revised International System of Units (SI), expected to be approved late in 2018, has implications for physics pedagogy. The ampere definition which dates from 1948 will be replaced by a definition that fixes the numerical value of the elementary charge, e, in coulombs. The kilogram definition which dates from 1889 will be replaced by a definition that fixes the numerical value of the Planck constant, h, in joule seconds. Existing SI equations will be completely unaffected. However, there will be a largely-negligible, but nevertheless necessary, change to published numerical factors relating SI electrical units to their corresponding units in the Gaussian and other CGS systems of units. The implications of the revised SI for electrical metrology are neatly illustrated by considering the interpretation of results obtained from a current balance in the present SI and in the revised SI. This manuscript was accepted for publication in February 2017 Online publication, April 2017. See: Richard S. Davis, American Journal of Physics 85, 364 (2017) http://dx.doi.org/10.1119/1.4976701 1 I. INTRODUCTION What is now known as the International System of Units (SI) adopted the ampere more than 60 years ago, preceded by lengthy discussions and a recommendation from the International Electrotechnical Commission (IEC).1,2 The definition of the ampere famously relies on the specification of a force per unit length of wire of a current balance that is manifestly impossible to construct. However, a simple current balance found in many undergraduate physics laboratories can be used to illustrate the ampere definition by measuring electric current and confirming that the fixed numerical value of the magnetic permeability of free space expressed in N/A2 leads to a result that agrees with that of a calibrated ammeter.3 Four of the seven base units of the SI are expected to be redefined in about two years’ time: the kilogram, ampere, kelvin and mole to create a system that is “more fundamental.”4,5 The first two of the new definitions will change how measurements with a current balance are interpreted, and these will be the focus of this article. The definitions of the remaining three base units of the present SI (the second, the meter, and the candela) will not be revised. The big picture is given in Refs. 4 and 5. The following analysis shows that the ideal current balance invoked in the present definition of the ampere will in the near future yield an experimental value for the fine-structure constant instead. In the following, the term “present” SI refers to the system of units defined in Ref. 1, which are in general use; the term “revised” SI refers to the revisions expected to be approved in the autumn of 2018.4,5 After its implementation, the revised SI will simply be “the SI.” 2 II. WHAT BECOMES OF THE CURRENT BALANCE IN THE REVISED SI? Ever since the SI was introduced, the ampere has been defined in terms of the measurement of a mechanical force. A version of the current balance can be found in many undergraduate physics laboratories and is used to measure an electric current in SI units, demonstrating how the present ampere definition can be realized.3,6,7 In the revised SI, the unit of charge, the coulomb, will be defined by giving a fixed (i.e. exact) numerical value to the elementary charge, e, when expressed in coulombs (by convention, the electron has charge e).4,5 The idealized current balance with infinitely long parallel wires, whose description appears in the present definition of the ampere, will no longer be mentioned. Rather, in the revised SI, the ampere will be realized from Ohm’s law by measuring the voltage U across a resistance R. The voltage will be measured in terms of the ac Josephson effect:4,5,8 h U n n , (1) 2e KJ where h is the Planck constant, is an adjustable microwave frequency, and n is an integral number of voltage steps. The Josephson constant, KJ, has the SI unit Hz/V. The resistance is measured using the quantum-Hall effect:4,5,9,10 11h RR, (2) i e2 i K where i is an integral number of impedance steps and RK, whose SI unit is , is known as the von Klitzing constant. The constants h and e appear in both equations. From Eqs. (1) and (2), the measured current I is U in Ie ν . (3) R 2 3 Since 1967, SI frequency measurements have been traceable to Cs , the hyperfine transition frequency of the cesium atom, whose numerical value in hertz has been fixed to define the SI second.1 In the revised SI, the numerical values of h and e will also be fixed to define the SI units J s (kg m2 s-1) and C (A s) respectively. Assigning a fixed numerical value to h also has the effect of redefining the kilogram because the second and meter are already defined.4 (In 1983 the numerical value of c, the speed of light in vacuum, was given a fixed numerical value in m/s to define the meter.11) The revised definition of the kilogram will no longer depend on the mass of a unique object manufactured in the 19th century, the international prototype of the kilogram (IPK).1,12 Continuity between the present and future definitions of the kilogram will be established by ensuring that the numerical value of h fixed in the revised SI equals the best experimental value of h in the present SI.13 The continuity between present and future definitions of the ampere is being established by the same strategy, as shown in detail below, using the current balance as an illustration. If the current balance will no longer exemplify the definition of the ampere as it does at present, what becomes of this metrological device and its simplified version used for laboratory demonstrations? We now answer this question. We begin with the form of Ampère’s force law used to define the ampere at present, I 2 F 2k , (4) L A a where FL is the force per unit length between two infinitely long parallel wires of negligible cross 1,14 section, separated by a distance a and maintained in vacuum. The constant kA is specific to the system of units that has been adopted, whereas the factor 2 is a geometric term applicable to the special case of the magnetic field produced by an infinitely long wire with current I flowing in it. Equation 4 shows the Lorentz force on a second wire, parallel to the first and carrying the same 4 current.15 The wires attract or repel depending on whether current in them flows in the same or opposite direction. The ampere definition specifies that if a = 1 m, a current of 1 A flowing in each 7 wire will result in FL having a magnitude of 2 10 N/m. In the SI, kA 0 / 4 . The present definition of the ampere is therefore equivalent to the specification 7 2 7 μπ0 / 4 10 N/A 10 H/m . The constant µ0 is usually referred to as the permeability of free space. These choices are the same as those of the rationalized MKS system of units, also known 1 2 14 as the MKSA system, which predates the SI. In the Gaussian-CGS system, kA 1/ c . Therefore electrical current in the Gaussian system has the unit dyn1/2 cm s1. The importance of Eq. (4) in what follows is its role in defining the ampere in the present SI, whereas the ampere will be defined differently in the revised SI. Although a continuity condition will ensure that the magnitude of the ampere will be essentially unchanged by the revision, there will be interesting consequences. In a typical device used to illustrate the Ampere force equation,3 the length L of each wire is finite but still reasonably long compared with a, i.e. L >> a. In that case the SI version of Eq. (4) can be approximated by F μI2 20 (5) L 4πa where F is a force determined by the weight mg of a mass m counter-acting the force between the wires, where g is the gravitational acceleration in the laboratory. Given that the relative uncertainty of I determined from these devices is typically of order 1 %, there is no need to introduce the correction of less than one part in 106 for the permeability of air relative to that of vacuum. Since 7 2 the definition of the ampere agreed in 1948, μπ0 /4 has had the exact value 10 N/A . In the revised SI, the factor 107 can no longer be exact because this would conflict with the decision to 5 fix the numerical value in coulombs of the elementary charge e. When the SI coulomb is redefined 2 in this way, the ampere is also redefined since 1 C = 1 A s. The numerical value of / 4 in N/A 0 must then be determined by experiment. Nevertheless, the value of e chosen will be consistent 7 2 with 0 / 4 10 N/A to within the smallest possible uncertainty at the time the revised SI is adopted. In fact, the relative uncertainty of µ0 will be identical to the relative uncertainty of the fine-structure constant, , which is a few parts in 1010 in the most recent recommendation.13 How did the fine-structure constant come into this picture? We can see by examining Eq.